Perplexity Enables Fully Local AI Workflows on Macs With 36GB RAM
Perplexity allows Pro, Max, and Enterprise users to run its Computer app entirely on Apple Silicon Macs, keeping data private.
Key points
- Macs with at least 36GB of RAM can now run Perplexity Computer workflows locally.
- Users can choose between Qwen 3.6 or PPLX Pro models for local execution.
- The feature is available to Pro, Max, and Enterprise subscribers starting today.
Perplexity has released an update allowing users to run its Computer app entirely on their own Macs, eliminating the need for cloud processing for certain tasks. This new capability requires Apple Silicon hardware with a minimum of 36GB of RAM, according to Engadget.
The update enables the entire Computer experience, including the harness and orchestrator, to operate locally. Perplexity previously introduced Hybrid Compute in September, which split workloads between frontier and local models. This latest release expands on that by permitting fully offline execution for eligible devices.
Local Processing and Model Options
Users can select from two specific models to run locally: Qwen 3.6 or PPLX Pro. PPLX Pro is a Perplexity-specific variant of Qwen 3.6 that underwent additional post-training to optimize it for Computer use. If a task requires more power, users can authorize the software to access more than 15 frontier models.
The company states that this feature benefits users concerned about data privacy and those looking to reduce credit spend. For instance, users can ask Computer to generate to-do lists from local Notes and Mail apps without sending email data to the cloud. Code and tool execution remain within an isolated sandbox environment.
Eligibility and Access
Perplexity Pro, Max, and Enterprise subscribers can begin using the local Computer feature immediately. To access it, users must download and install the latest version of the Perplexity Mac app. The software provides in-app instructions for downloading and installing the chosen local model on the machine.
Perplexity notes that as demanding workflows become practical for local execution, personal devices will handle a larger share of AI work. This shift gives users greater control over sensitive data and computing costs. The security protocols used in the local sandbox mirror those in the cloud version.
What should you do?
Users interested in this feature should verify their Mac meets the 36GB RAM requirement and runs on Apple Silicon. Subscribers to Pro, Max, or Enterprise plans should update the Perplexity Mac app to the latest version. Those concerned with privacy can review the in-app instructions for installing local models.
What to do and how to stay safe: Perplexity
- Verify your Mac has Apple Silicon and at least 36GB of RAM before attempting to use local features.
- Update the Perplexity Mac app to the latest version to access the new local workflow options.
- Review the in-app instructions carefully to download and install the preferred local model correctly.
- Monitor system performance when running local models, as they may consume significant resources.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What are the system requirements for running Perplexity Computer locally?
You need a Mac with Apple Silicon and at least 36GB of RAM to run the workflows entirely on your device.
Which AI models can be run locally?
Users can choose between Qwen 3.6 and PPLX Pro, a Perplexity-fine-tuned version of Qwen 3.6.
Who can access this feature?
Perplexity Pro, Max, and Enterprise subscribers can use the local Computer feature starting today.
